PROJECT SUMMARY Despite a broad understanding of the multi-level determinants of health disparities in childhood obesity, known risk factors including diet and physical often do not adequately predict later obesity among low-income and minority children. Developing effective intervention targets for childhood obesity in these populations is therefore difficult. Emerging evidence suggests that ultra-processed food consumption may partially explain health disparities in cardiovascular disease and cancer among adults. However, existing methods to classify foods based on the extent of processing are inconsistent and unclear, making it difficult to assess exposure to ultra-processed foods and associations with childhood weight outcomes. The current proposal will apply the NOVA classification for ultra-processed foods to dietary recall data from an NHLBI-funded childhood obesity prevention randomized controlled trial. Developing a reliable and valid methodology for assessing ultra- processed food consumption and testing associations with childhood obesity will generate evidence to better characterize dietary intake and to identify potential targets for reducing health disparities in childhood obesity. This proposal builds on a robust dataset from the Growing Right Onto Wellness (GROW) trial, which aimed to prevent childhood obesity using a three-year multi-level and culturally-tailored intervention. The trial randomized 610 parent-preschool child pairs and achieved >90% retention at three-year follow-up, with high rates of data completeness. Eligible children were ages 3-5 at enrollment, spoke English or Spanish, and had BMI ?50th percentile and <95th percentile. The dataset includes 24-hour diet recall data collected using NDS-R software at baseline and three annual follow-up timepoints. The current proposal will develop and validate a novel coding algorithm to map the existing NOVA classification system for ultra-processed foods onto this diet recall data. This algorithm will generate an analytic variable that describes the number of calories consumed per day in each of the four NOVA classifications for food processing level. Using the newly developed approach to assessing ultra-processed food consumption, we will test the association between higher levels of ultra-processed foods and childhood obesity across 3 years of follow up. The main exposure variable will be the number of daily calories consumed for ultra-processed foods and the primary outcome will be child raw BMI. The goals of this proposal are to 1) advance dietary measurement by developing a novel methodology for evaluating levels of ultra-processed food consumption using diet recall data; 2) assess whether ultra-processed food consumption level is predictive of incident obesity among low- income, minority preschoolers; and 3) develop evidence for intervention targets for future R01 funding.